Risks in Social Media and Deep Learning

Social media is pervasive and intensely scrutinized by deep learning systems. It started as a way to gather "eyeballs", to sell ad space. But nobody paid enough attention for branding campaigns to be very effective, let alone product information spots. Despite the inability for traditional advertising to "stick", deep learning and individual targeting showed that social media is surprisingly effective at herding people to buy trendy things. While this ability to influence crowd behavior has been useful in commerce, a new and far more lucrative application is taking over: Changing the course of nations. And nothing influences group behavior so cost efficiently as fostering petty emotions like fear and hate. Robert's talk will explore the theory, recent applications, and potential destinations of this shift. Then he will present some of the countermeasures for immediate personal defense and longer-term solutions for protecting our society and the ability to think freely.